PURPOSE: To assess the additional damage of normal tissues attributable to reirradiation and the magnitude of partial recovery following the initial course. METHODS AND MATERIALS: Symptomatic late complication rates (excluding xerostomia) in 3635 patients receiving one course (Group 1) and 487 patients receiving two courses of external radiotherapy (Group 2) for nasopharyngeal carcinoma were retrospectively analyzed and compared. RESULTS: Group 2 had significantly lower actuarial complication-free survival rates than Group 1: 48% versus 81% at 5 years. The post-retreatment incidence was significantly affected by biologically effective dose (BED) (assuming an alpha/beta ratio of 3 Gy) of the first course: hazard ratio (HR) = 1.04 per Gy(3) (p = 0.01), but only marginally by that of the second course: HR = 1.01 per Gy(3) (p = 0.06). If the summated BED was taken as the dose unit, it was estimated that a total BED of 143 Gy(3) would induce a 20% incidence at 5 years, while the corresponding dose projected from Group 1 was 111 Gy(3). The gap effect was insignificant in the overall analyses, but a trend of decreasing risk with increasing interval was observed in patients with gap > or = 2 years: HR = 0.86 per year (p = 0.07). CONCLUSION: The major determinant of post-retreatment complication is the severity of damage during the initial course. The sum of total doses tolerated is higher than that expected with a single-course treatment, suggesting occurrence of partial recovery (particularly in those reirradiated after an interval of 2 years or more).
PURPOSE: To assess the additional damage of normal tissues attributable to reirradiation and the magnitude of partial recovery following the initial course. METHODS AND MATERIALS: Symptomatic late complication rates (excluding xerostomia) in 3635 patients receiving one course (Group 1) and 487 patients receiving two courses of external radiotherapy (Group 2) for nasopharyngeal carcinoma were retrospectively analyzed and compared. RESULTS: Group 2 had significantly lower actuarial complication-free survival rates than Group 1: 48% versus 81% at 5 years. The post-retreatment incidence was significantly affected by biologically effective dose (BED) (assuming an alpha/beta ratio of 3 Gy) of the first course: hazard ratio (HR) = 1.04 per Gy(3) (p = 0.01), but only marginally by that of the second course: HR = 1.01 per Gy(3) (p = 0.06). If the summated BED was taken as the dose unit, it was estimated that a total BED of 143 Gy(3) would induce a 20% incidence at 5 years, while the corresponding dose projected from Group 1 was 111 Gy(3). The gap effect was insignificant in the overall analyses, but a trend of decreasing risk with increasing interval was observed in patients with gap > or = 2 years: HR = 0.86 per year (p = 0.07). CONCLUSION: The major determinant of post-retreatment complication is the severity of damage during the initial course. The sum of total doses tolerated is higher than that expected with a single-course treatment, suggesting occurrence of partial recovery (particularly in those reirradiated after an interval of 2 years or more).